Wednesday, May 13, 2009

On Graham Harrell, "system quarterbacks" and the NFL

This past season in American college football, Texas Tech quarterback Graham Harrell broke the NCAA record for most touchdown passes thrown in a college career with 134. He completed 442 of 626 passing attempts for an outstanding 70.6 completion percentage, threw for 5,111 yards, completed 45 touchdown passes while only being picked off nine times and became the first NCAA quarterback to record multiple seasons with more than 5,000 yards passing. Despite all that, Harrell wasn't selected in the NFL draft. In the first five picks alone, the Detroit Lions and New York Jets opted to take Matthew Stafford and Mark Sanchez respectively, despite each accomplishing significantly less than Harrell at the college level. By comparison, Stafford completed only 235 passes on 383 attempts at Georgia last season for 3459 yards and a completion percentage of 61.4 per cent; he also threw 25 touchdowns and 10 interceptions. Sanchez completed 241 of 366 attempts (65.8 per cent) at USC for 3207 yards with 34 touchdowns and 10 interceptions. Moreover, both Sanchez and Stafford didn't have a great track record previously; Sanchez wasn't USC's top quarterback in the previous year and Stafford hadn't cracked 3000 yards in two seasons at Georgia. By comparison, Harrell put up 5705 yards in 2007 with 48 touchdowns, 14 interceptions and a 71.8 per cent completion percentage. Yet, in the draft, NFL teams concluded that Stafford and Sanchez were the first- and fifth-most valuable players respectively, while Harrell didn't even crack the top 256. It didn't get much better afterwards; Harrell wound up attending the Cleveland Browns' training camp, but many observers of the team, including Tony Grossi of the Cleveland Plain Dealer see him as a long shot to even make the practice squad.

Why was Harrell so overlooked? The curse of the dreaded label; "system quarterback". It's a bizarre term, as by definition, all quarterbacks play in one system or another. Yet, "system quarterback" is a pejorative and derogatory term hurled at those college quarterbacks who put up great numbers in pass-oriented offences like Texas Tech's, many of which prominently feature the spread. System quarterbacks tend to be seen as a product of their environment, doomed to failure outside of college ball.

To be fair, there is some historical evidence supporting this idea many NFL teams seem to have. Andre Ware, Gino Torretta and Eric Crouch all won the Heisman Trophy in college but didn't make much of an impact in the NFL. Of course, some of that was due to a lack of opportunity; Crouch didn't even get much of a chance to play quarterback in the NFL due to his height. Still, there is precedent of great college quarterbacks failing to adjust to the NFL, and many of them came from pass-heavy offences.

However, who will make a good NFL quarterback is one of the most difficult things to predict, as the great Malcolm Gladwell relates in Outliers. Teams have tried everything from college stats to height requirements to arm strength to Wonderlic tests, but still haven't found a consistent way to pick who will be a good quarterback at the NFL level. Lately, the focus has been on tall, athletic players with good mobility and arm strength who have played in offences similar to the traditional ones used in the NFL, but even that hasn't always panned out; see JaMarcus Russell, Alex Smith, Jason Campbell and Vince Young for examples.

So, if both groups of players frequently become busts, why are players with lesser stats but more "desirable" physical attributes still so favoured by NFL teams? For one thing, there's the lure of potential; you can always imagine a guy with a rocket arm learning to make good reads and precise throws under pressure, but it's hard to picture an intelligent, high-percentage quarterback with a weak arm suddenly becoming able to throw downfield bombs. That doesn't mean it's true, though. It's like the equivalent of baseball's plate discipline. As Michael Lewis detailed in Moneyball, Billy Beane was able to get unathletic guys who put up great on-base percentages in high school or college rather cheaply because most of the other teams were more concerned with players' swing mechanics and speed. They figured they'd win by taking great athletic specimens, and in many cases, it worked; however, everyone was going in the same direction, so those players became quite hard to find. Moreover, Lewis relates that many of those teams figured that players with low on-base percentages would improve them over time as they learned discipline; plenty of research has shown that players don't tend to change their hitting approach, though. It's a similar story to the NFL; it's easier to imagine someone gaining plate discipline or route-reading abilities than tremendous speed or a rocket arm, but that doesn't mean it's much more likely to happen. Some batters will always go up to the plate hacking away with tremendous swing mechanics, and some quarterbacks will always be gunslingers who have great arms but are often picked off.

Another aspect of this bias is because of highlights. Sports tend to be all about highlights these days, and much of the NFL scouting process is based on looking at film of players. It's always going to be more impressive to watch a quarterback successfully chuck a 40-yard bomb to a receiver in double coverage than a 8-yard quick out to an open receiver near the sidelines, even if the latter is a much better play nine times out of 10. The high-completion percentage, low-interception quarterbacks tend to excel via great route-reading skills, an ability to think on their feet and a knack for safely getting the ball to an open man, all of which take a tremendous amount of talent but don't easily translate into highlight packages. Meanwhile, the strong-armed daredevils make spectacular throws that could just as easily turn into interceptions as touchdowns, but look better on film. It's why the New York Jets dumped notoriously weak-armed Chad Pennington for aging gunslinger Brett Favre before last season; Pennington wound up with the Dolphins and put up great stats with them, while Favre self-destructed down the stretch with a ton of interceptions. Pennington finished the year with 19 touchdowns against seven interceptions, a 67.4 completion percentage, 3,653 passing yards and a 97.3 passer rating, while Favre finished with 22 touchdowns, 22 interceptions, a 65.7 percentage, 3,472 yards and a 81.0 rating. Pennington was superior in every category except touchdowns, but Favre wound up on the highlights more frequently thanks to his habit of forcing the ball into dangerous situations that either resulted in a spectacular catch or an interception.

The final element to consider is the NFL systems involved. If you ever catch NFL analysts talking about quarterbacks, something that invariably comes up is their ability to make "The NFL Throws". This has been one of the biggest knocks against quarterbacks like Harrell that aren't considered to have the arm strength to throw 30- to 40-yard bombs. However, this presumes that those throws are necessary for success in the NFL, when in reality, they're generally the least likely to succeed. That's not to say there's never an occasion where a Hail Mary is required, but rather to suggest that such occasions are few and far between and success on them perhaps isn't necessarily the best benchmark for NFL quarterback success.

Part of the problem is that there's a certain element of Orwellian groupthink involved in the NFL. NFL head coaches are frequently former players and almost always guys who have been around the league in assistant capacities for a long time. There isn't a lot of fresh blood or original thinking, and as a result, most teams' systems and offensive schemes tend to be somewhat similar. Sure, there's plenty of variations on the different themes involved, but the general plan of an NFL franchise's offence involves a big and strong quarterback who can throw bullet passes, a bruising running back who can crash up the middle for "three yards and a cloud of dust" each down, a couple of tall and lightning-fast wide receivers who can streak downfield and perhaps a pass-catching tight end or slot receiver for the occasional shorter throw.

The big flaw with this schematic is it puts substantial limitations on the types of players you can draft. There are only a certain number of quarterbacks who fit this mould, only a certain number of wide receivers with the speed you need and pass-catching skills to with them, and only a certain amount of running backs who will match your physical specifications. Like baseball in the old days as well, this is excarbated by every other team also looking for the same types of players. There's a large demand and a limited supply, which drives the price up. If you get lucky with draft position or free-agent signings, you might be able to get some of the players you need, but it won't be easy.

How do you get around this? It's not particularly simple. The reason these ideas and schemes have persisted for so long is because they do tend to work. The basic offence described above is reasonably well-balanced and ideally is full of athletic players who can execute their roles to perfection. Each player is generally strong in several areas, such as running backs with speed and size and receivers with speed and hands. As previously mentioned, this makes it difficult to build an entire roster of these types due to the demand and the resulting costs, but the basic idea isn't a bad one.

You can get a little creative, though. If you'll permit me a little geekery here, consider football as a roleplaying game (say, a Dungeons and Dragons-based one) for a moment. These games work by allowing you to pick a race and a class, each with their own unique abilities. You can then further customize your character with specific attributes, skills, feats and equipment. However, there is a cost involved; you only have a certain number of points or resources you can allot to each area, and your class and race helps in some areas but weakens others. Moreover, you can pick up some skills and such from other classes along the road, but they're more expensive to develop. Thus, the real path to success is by picking one or two areas that you're going to excel in and focusing on developing them. You can have a fighter who shines in close combat, a ranger who prefers to snipe from afar, a wizard with strong magical attacks or any variety of other types, but it's almost impossible to create a character that's skilled in all areas. If you attempt to be good at everything, you excel at nothing. Instead, it works much better to star in one area and count on the other characters in your party to take care of your weaknesses.

This can be applied to football quite successfully. It's very rare to find those players that are good at everything and have all the physical attributes you want at a position. When they do show up, they're so expensive that it's incredibly tough to land them. More frequently, you have to make tradeoffs, which explains why Detroit and New York chose Stafford and Sanchez in the draft; they had some of the athletic attributes they were looking for and they're hoping that they can develop the route-reading skills necessary for success. They may fit the traditional NFL system or they may not; we'll have to see.

However, Stafford and Sanchez did shine in many of the areas highly regarded by scouts and they have some skills in all areas, which of course made them rather in demand and thus expensive. Trying to build a franchise according to this model requires a succession of high draft picks or expensive free agency signings, and there isn't much room for error because of the cost of each area.

The alternative is to take a Moneyball-esque approach and target talented players who are undervalued, such as Harrell. Now, these players tend to be undervalued because they have significant flaws in the eyes of the typical NFL model. They usually aren't as versatile as you'd like and they won't be stars in a traditional NFL system. The advantage of this is they tend to make up for their flaws in other ways. To revert to the D&D example, it isn't a big problem for a wizard to not be the best hand-to-hand fighter thanks to their magical skills. The other benefit of these players is they tend to come at a considerably lower cost, making the price of failure much more acceptable.

What you then need to do is come up with a system that can utilize these players effectively. The classic example of this is legendary San Francisco head coach Bill Walsh's West Coast offence. Walsh started with a cast of players no one else was that excited about and made a key move when he drafted quarterback Joe Montana 82nd overall in 1979. Montana's resume to that point reads much like Harrell's; he had a great college career at Notre Dame, but was overlooked by most scouts thanks to his unexceptional arm strength. What he did have going for him was tremendous accuracy, which Walsh made great use of in a new offensive system that emphasized tons of quick and accurate short passes instead of the traditional strategy pounding the ground with bruising runs up the middle and then throwing bombs downfield. Under Montana and Steve Young, a quarterback in a very similar mould, the 49ers were one of the dominant teams of the 1980s and claimed five Super Bowls between 1981 and 1994.

That's not to say that the West Coast offence is the entire answer; in fact, it's been quite widely adopted since them and many NFL teams now use at least some elements of the strategy. The key is Walsh's overarching philosophy of finding players that were overlooked by other teams due to some weaknesses and then designing schemes to take advantage of their strengths. There's no divine commandment that mandates NFL teams to play a certain way, but most of them adopt similar strategies due to conservatism; you aren't often questioned if you follow the crowd.

It's largely the same in college ball, but some underdog teams and coaches are more willing to try unusual strategies. Mike Leach, Harrell's coach at Texas Tech, is one such unconventional thinker (and surprise, surprise, he's been profiled twice by Michael Lewis). Leach couldn't compete with the big schools like Texas and Oklahoma in recruiting in-demand, all-around athletes who fit the profiles everyone else was using, so he developed his own pass-wacky system using undervalued quarterbacks who were extremely accurate but didn't have much else going for them. Here's the money quote from Lewis' first profile of Leach:

"[Current Detroit Lions head coach Jim] Schwartz had an N.F.L. coach's perspective on talent, and from his point of view, the players Leach was using to rack up points and yards were no talent at all. None of them had been identified by N.F.L. scouts or even college recruiters as first-rate material. Coming out of high school, most of them had only one or two offers from midrange schools."

These aren't the all-around players used by most schools, but Leach found a way to mould them into a tremendously successful team. That suggests that there's more ways to win than traditionally thought by many football coaches and analysts. We're seeing this more in the NFL as well; look at the Dolphins' success with the Wildcat offence last year, which prompted them to draft Pat White this year. White's another quarterback who likely wouldn't have had much of a chance in the NFL in a traditional mould, but has a chance to shine in the Wildcat. Like White, Harrell isn't the perfect fit for the standard NFL offence, but that doesn't diminish his talent; all he needs is the right system to succeed. The NFL's too concerned with the arms race for all-around players at the moment and often overlooks those who don't quite meet its physical expectations. The question is if there's a Billy Beane or Mike Leach out there who can exploit that.

One final thought on Harrell; if he isn't able to catch on with an NFL team this year, might he come north of the border? CFL offences are already heavily pass-based, and some of the multiple receiver sets are pretty close to what Harrell was used to at Texas Tech. Plus, the CFL has a long and detailed history of innovation, so perhaps a Texas Tech-style offence is the next natural step. The CFL's also been an excellent proving ground for those overlooked by the NFL in the past, such as Doug Flutie and Warren Moon. Could Harrell be the next great quarterback to follow in their footsteps? We'll never know unless someone gives him a shot.


  1. Anonymous4:03 AM

    I hope Harrell goes to the CFL. I'd like to see him succeed and have fun there. That kid is a champion!

    This was a great article!

    Leach is awesome! He leads the Big 12, and is somewhere in the top 25 in the nation for the category that matters most: Player Graduations.

    Guns Up, Red Raiders!

  2. Anonymous10:54 PM

    Outstanding article! Hopefully Graham is currently en route to Rider Nation as we speak!

  3. Anonymous6:01 PM

    well he has been sighed by the saskatchewan roughriders. this club won the grey cup in 2007 and had a great year in 2008 dispite having more injuries than any other club in the league. this year they are 2-0 and odds at 13-5 to be in the final. the orgaization is sound and they have, hands down, the best fans in the league. this could be a fantastic experiance for him and a spring board to greatness in the nfl.i'm hoping he can bring another championship to my riders.